One of our reviewers reports that when they clicked save after doing a review, i disappeared.
I checked the log for the paper and no record of the review is there. Any ideas what could be causing this? Ideally, if there is some way to recover the review, that would be better still.
Hi @philipmach,
Can you describe the behavior more closely? Did the reviewer appear to have been logged out by the system, or were they returned to the review page to no apparent sign of the review, or something else?
Regards,
Alec Smecher
Public Knowledge Project Team
As described by the reviewer (who is also watching this and may add detail if needed):
I saved regularly while I was busy, and after I went back to the ‘front’ page to finalise, it would not allow me the final recommendation because I had not ‘submitted a review’, and everything was gone
Hi @philipmach,
Are you using one-click reviewer access? See Journal Setup, step 2.
Is the session_check_ip setting enabled in config.inc.php?
Regards,
Alec Smecher
Public Knowledge Project Team
Yes to the first question; I will have to find out from our sysadmin on the second since as far as I know the PHP layer is not exposed to the user interface.
My sysadmin says session_check_ip is set to off.
Hi @philipmach,
Hmm, none of the usual suspects seem to be active in your situation.
I double-checked a few potential causes that might result in the review form being saved blank – missing POST data, for example, which (if missing) might result in unexpected behavior when using the “Back” button – and it looks like the necessary checks are in place.
On some servers that use mod_security or equivalent, certain words in certain orders can get incorrectly flagged as attempts to break into the server. These are generally logged. This might result in a valid POST request getting chopped in half without OJS being informed, which could result in lost data. If this is the case, then you should see something logged in your server’s security log for mod_security (which might be in a variety of places).
Unfortunately, in any case the review is very likely gone. Web servers don’t typically log the details of POST requests, and OJS seems to have received a request to save the review with no contents in the fields.
This is the only report I can recall of this issue, which suggests to me an unusual failure – a network hiccough, browser crash, etc. – rather than a misconfiguration. The rarity of these happening and – frequently – combination of infrequent factors make these terribly difficult to track down.
I don’t believe anything relevant has changed between 2.4.5 and 2.4.8-1, but it’s always a good idea to stay up to date with stability releases – there are a variety of minor bugfixes and tweaks included since 2.4.5 was released.
Regards,
Alec Smecher
Public Knowledge Project Team
Thanks – it probably is something rare even if related to our install because I don’t recall seeing it with any other review.